{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:KK5NMMBETRUMDRMIIENHHVYVMT","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"b36989afa6529509c7307cc2d34d6449f27efaad915c71ce25b25673e59c75dc","cross_cats_sorted":["cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"stat.ML","submitted_at":"2024-10-10T19:06:23Z","title_canon_sha256":"30a814eb38afebaf66b1f85cc50fcd3715d0d317580cd3f100d4e66b54f1d3fa"},"schema_version":"1.0","source":{"id":"2410.08315","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2410.08315","created_at":"2026-07-05T09:18:56Z"},{"alias_kind":"arxiv_version","alias_value":"2410.08315v1","created_at":"2026-07-05T09:18:56Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.08315","created_at":"2026-07-05T09:18:56Z"},{"alias_kind":"pith_short_12","alias_value":"KK5NMMBETRUM","created_at":"2026-07-05T09:18:56Z"},{"alias_kind":"pith_short_16","alias_value":"KK5NMMBETRUMDRMI","created_at":"2026-07-05T09:18:56Z"},{"alias_kind":"pith_short_8","alias_value":"KK5NMMBE","created_at":"2026-07-05T09:18:56Z"}],"graph_snapshots":[{"event_id":"sha256:41638168358880f23c00083367d70646d9b43020d6dd96fe17d0d3b38d5593c3","target":"graph","created_at":"2026-07-05T09:18:56Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2410.08315/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Fine-tuning foundation models via reinforcement learning (RL) has proven promising for aligning to downstream objectives. In the case of diffusion models (DMs), though RL training improves alignment from early timesteps, critical issues such as training instability and mode collapse arise. We address these drawbacks by exploiting the hierarchical nature of DMs: we train them dynamically at each epoch with a tailored RL method, allowing for continual evaluation and step-by-step refinement of the model performance (or alignment). Furthermore, we find that not every denoising step needs to be fin","authors_text":"Crist\\'obal Alc\\'azar, Felipe Tobar, Roberto Barcel\\'o","cross_cats":["cs.LG"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"stat.ML","submitted_at":"2024-10-10T19:06:23Z","title":"Avoiding mode collapse in diffusion models fine-tuned with reinforcement learning"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.08315","kind":"arxiv","version":1},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:1ed561e5f9cfbd2f58221eaab0f247f31701e095f4a03bd9df66532f2178f42e","target":"record","created_at":"2026-07-05T09:18:56Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"b36989afa6529509c7307cc2d34d6449f27efaad915c71ce25b25673e59c75dc","cross_cats_sorted":["cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"stat.ML","submitted_at":"2024-10-10T19:06:23Z","title_canon_sha256":"30a814eb38afebaf66b1f85cc50fcd3715d0d317580cd3f100d4e66b54f1d3fa"},"schema_version":"1.0","source":{"id":"2410.08315","kind":"arxiv","version":1}},"canonical_sha256":"52bad630249c68c1c588411a73d71564f74c861e32f8995d66c6e55e707212b5","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"52bad630249c68c1c588411a73d71564f74c861e32f8995d66c6e55e707212b5","first_computed_at":"2026-07-05T09:18:56.129555Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T09:18:56.129555Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"LXnsC9zIdZFZYK9YZo9VBRFT3hssUGyS3YnOuGP7qOeCZN++pcNWmckWnAXfNRCIlNw4VQk2/r/K/srdsNdEBg==","signature_status":"signed_v1","signed_at":"2026-07-05T09:18:56.129975Z","signed_message":"canonical_sha256_bytes"},"source_id":"2410.08315","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:1ed561e5f9cfbd2f58221eaab0f247f31701e095f4a03bd9df66532f2178f42e","sha256:41638168358880f23c00083367d70646d9b43020d6dd96fe17d0d3b38d5593c3"],"state_sha256":"f5fc606e14a5eac110e01a549e3182c1c2f0454b6773945d29d664a01c635637"}